US20150331006A1

US20150331006A1 - Velocity Measurement Device and Velocity Measurement Method

Info

Publication number: US20150331006A1
Application number: US14/648,826
Authority: US
Inventors: Jyunichi Urata; Masayuki INABA
Original assignee: Schaft Inc
Current assignee: Schaft Inc; Google LLC
Priority date: 2012-12-26
Filing date: 2013-12-25
Publication date: 2015-11-19
Also published as: WO2014103931A1; EP2940479A1; JPWO2014106937A1; WO2014106937A1; JP6325458B2; EP2940479A4; EP2940479B1; CN105008934A

Abstract

The present invention aims to easily suppress noise components of aliases when measuring a velocity based on an output from an encoder. The invention comprises a step for converting in which two-phase signals indicating the movement of a spinning body, etc., are converted to a ternary pulse train signal having a positive or a negative or a zero value corresponding to the phases of the two-phase signals, a step for low pass filtering in which a low frequency component of the pulse train signal is obtained by arithmetically processing the pulse train signal and filter coefficients, and a step for output-sampling in which the signal obtained by the step for low pass filtering is sampled at a predetermined sampling period. In the step for low pass filtering, a convolution (operation for convolution) between the pulse train signal obtained by sampling the pulse train signal at a predetermined period and the filter coefficients is carried out. Sampling the pulse train signal is synchronized with and has the same phase as reading the filter coefficients for the convolution. The sampled signal by the step for output-sampling is output as a velocity signal corresponding to the velocity determined by the two-phase signal.

Description

TECHNICAL FIELD

The present invention relates generally to measuring a velocity, and more particularly to measuring a velocity based on an output from an encoder.

BACKGROUND OF THE INVENTION

For measuring a revolving speed of a spinning body such as a revolving shaft, using an encoder such as a rotary encoder that outputs a signal synchronized with the revolution of the body is well-known. For measuring a moving velocity of a linearly moving body, using a linear encoder that outputs a signal synchronized with the movement of the body is well-known.
FIGS. 9 and 10 are explanatory drawings showing an overall configuration of a conventional device 101 for measuring a velocity. By way of example, a device for measuring a revolving speed of a spinning body is now explained. In FIG. 9, an encoder 100 outputs an output signal synchronized with the revolution of a body such as a revolving shaft. For example, an incremental encoder outputs two-phase output signals, i.e., an A-phase signal and a B-phase signal (FIGS. 10 a and 10 b). The relationship between their phases differs depending on the direction of the revolution. Thus, the direction of the revolution can be detected from the relationship between the phases of the two-phase signals.
A counting part 102 counts the signals output from the encoder. Counting the signals can be carried out by counting the pulses in the pulse train signal (FIG. 10 c) that are obtained at the edges of the signals output from the encoder by a counter. FIG. 10 d shows a counted value.
A sampling part 103 samples counted values output from the counting part 102 at predetermined time intervals. A part 104 for detecting differentials outputs the temporal changing rate of the sampled values (shown by circles in FIG. 10 d) as a velocity (FIG. 10 e).
Since a value obtained by counting pulses in a pulse train signal in a predetermined period corresponds to the frequency of the signal, measurement of such a frequency by using such a value is well-known. The information about the frequency of the pulse train signal is in the low frequency component in the frequency spectrum of the value. Therefore, it is well-known to pick out the low frequency component by eliminating the higher harmonic component caused by a quantization error by using a low pass filter (Patent Literature 1; Japanese Patent Publication of Examined Application No. 2009-250807).

SUMMARY OF THE INVENTION

A conventional method generates noises such as quantization errors when a signal output from an encoder is sampled. If a velocity is measured based on the signal that includes the noises, it cannot be measured with a high accuracy because the signal is caused to fluctuate by the noise components. The quantization noise can be eliminated by a low pass filter, because the frequency of the noise is higher than the frequency (fi) of the pulse train signal output from the encoder.
However, an alias-component, which is generated by sampling, exists in a low frequency band. Accordingly, the component cannot be eliminated by the low pass filter. FIG. 10 f shows the relationship among the frequency (f_i) of the signal to be encoded, the sampling frequency (f_s), and the frequency (f_s-f_i) of the alias.
Since the frequency (f_s-f_i) of the aliases is less than the frequency (f_i) of the signal to be encoded, a low pass filter (LPF) cannot eliminate the frequency component of the alias that has the frequency (f_s-f_i). The dashed line in FIG. 10 f shows the characteristics of the low pass filter (LPF).
The invention of this application aims to resolve the above problem, i.e., to easily suppress the components of a quantization noise and an aliasing noise when a velocity is measured based on an output from an encoder.
The invention of this application does not carry out a conventional process in which the values got by counting the pulse train signal are sampled and then the signal composed of the sampled values is passed through a low pass filter. But instead it converts a two-phase encoded signal that indicates a movement of a spinning body to a ternary pulse train signal. Then, it directly causes the ternary pulse train signal to pass through a low pass filter, and obtains a velocity signal by sampling the signal having been passed through the filter.
This invention, for measuring a velocity, causes a pulse train signal to directly pass through a low pass filter and thus reduces the high frequency components of the pulse train signal. Thus, it can reduce the alias components induced by the subsequent process for sampling the signal.
The device for measuring a velocity of this invention comprises a part for pulse-converting that converts two-phase signals to a ternary pulse train signal having a positive or a negative or a zero value corresponding to the phases of the two-phase signals, a low pass filter that outputs a low frequency component of the pulse train signal by means of arithmetical processing between the ternary pulse train signal and filter coefficients, and a part for output-sampling that samples the signal output from the low pass filter at a predetermined sampling period and outputs a velocity signal corresponding to the velocity determined by the two-phase signal.
The two-phase signal is an encoded pulse train obtained by an encoder. The relationship between the phases of the two-phase signals is the following: one of the phases is advanced or delayed in relation to the other, depending on the direction of a movement of an object such as the direction of a revolution. From the relationship, the information on the direction of the revolution, etc., can be obtained.
The part for pulse-converting converts two-phase signals to an up and down pulse train signal having a positive or a negative or a zero value corresponding to the relationship between the phases of the two-phase signals. The pulse train signal can be made by detecting the edge portions of the two-phase signals. A positive or negative sign can be determined by the change of the relationship between the phases of the two-phase signals.
The low pass filter is a digital filter. The filter reduces the high frequency components in a pulse train signal, allows low frequency components in the signal to pass through, and obtains velocity signal components, by using an arithmetical processing that carries out a convolution (operation for convolution) between the pulse train signal and the filter coefficients. The part for output-sampling samples the signal output from the low pass filter at a predetermined sampling period. The process for causing low frequency components to pass through the low pass filter and the process for sampling at the part for sampling result in outputting a velocity signal corresponding to the velocity determined by the two-phase signal.
By the process for sampling, measuring a velocity is carried out. That velocity corresponds to the pulse density of the pulse train signal within a certain sampling period. The sampling period can be arbitrarily set depending on the pulse density. For example, if the pulse density of the pulse train signal is high and the intervals between the pulses in a pulse train signal are short, the pulse train signal may sharply change in a short time. In such a case, the sampling period should be set to be short. In contrast, if the pulse density of the pulse train signal is low and the intervals between signals in a pulse train signal are long, the pulse train signal cannot sharply change in a short time. In such a case, the sampling period can be set to be long.

(An Embodiment of the Low Pass Filter)

The low pass filter carries out an arithmetical process that picks out the low frequency component by eliminating the high frequency component. The process carries out a convolution (operation for convolution) between the pulse train signal and the filter coefficients. The filter coefficients can be, say, coefficients in a FIR filter.
The low pass filter comprises a counting part, a part for input-sampling, a multiplexer, an adder and a subtracter, a part for storing filter coefficients, and a part for storing the results. The counting part outputs timing signals. The part for input-sampling samples a pulse train signal. The multiplexer carries out a process for selecting signals. The part for storing filter coefficients outputs filter coefficients that are to be used for arithmetical processing in the adder and the subtracter. The part for storing the results stores the results of the convolution between the pulse train signal or the result of the arithmetical processing and the filter coefficients.
The counting part outputs a readout signal, an input-sampling signal, and an output-sampling signal having a sampling period, as a timing signal. The readout signal has a predetermined cycle. The input-sampling signal is synchronized with and has the same phase as the readout signal. The timing signal is used for arithmetical processing, etc. Also, the counting part can output a reset signal. The reset signal has a sampling period synchronized with the output-sampling signal.
The part for input-sampling samples the pulse train signal based on the input-sampling signal, and inputs the sampled signal into the multiplexer.
The part for storing filter coefficients stores filter coefficients that have been previously set. The part reads out the filter coefficients when the part receives the readout signal and then outputs the coefficients to the adder and the subtracter, so as to set the filter coefficients for the low pass filter.
The adder adds the filter coefficients that are read out from the part for storing the filter coefficients to the signal output from the part for storing the results. The subtracter subtracts the filter coefficients, which are read out from the part for storing the filter coefficients, from the signal output from the part for storing the results.
The multiplexer receives the pulse train signal as a control signal for selection. Based on the signal, it outputs one of the two signals from the adder and the subtracter, or outputs the same signal that is output from the part for storing the results.
When the pulse train signal is positive, the multiplexer selects and outputs the signal output from the adder, among the signals input to the multiplexer, i.e., among the signal output from the subtracter, the signal output from the subtracter, and the signal output from the part for storing the results. When the pulse train signal is negative, it selects and outputs the signal output from the subtracter. When the pulse train signal is zero, it selects and outputs the signal output from the part for storing the results. This process is carried out under the condition that the input-sampling signal is synchronized with the readout signal, and repeated while the filter coefficients are being read out in sequence. Thus, a convolution (operation for convolution) between the pulse train signal and the filter coefficients is carried out.
When a FIR filter is used, the part for storing the results can carry out the convolution (operation for convolution) for every sampling period, by resetting the result of the convolution at every sampling period by using a reset signal synchronized with the output-sampling signal.
If a FIR filter (finite impulse response filter) is configured as a low pass filter, the counting part includes a reset signal, as a timing signal, for resetting a sampling period. An adder adds filter coefficients read out from a part for storing filter coefficients to a signal output from a part for storing the results. A subtracter subtracts filter coefficients read out from the part for storing filter coefficients from a signal output from the part for storing the results. A part for storing the results resets the result of the convolution at every sampling period by using a reset signal.
A part for output-sampling samples a signal output from the multiplexer at every sampling period at the time of the output-sampling signal, and outputs the sampled signal as a velocity signal.
By the low pass filter of the invention of this application, filter coefficients that are used for a convolution at the low pass filter are stored in a part for storing filter coefficients. Therefore, the coefficients can be easily changed and the change of the characteristics of the filter can easily be made.
The low pass filter of the invention of this application is configured to combine an adder, a subtracter, and a multiplexer. By inputting filter coefficients to the adder or the subtracter, a convolution between a pulse train signal and the filter coefficients can be carried out with no multiplier.
The device for measuring a velocity of the invention of this application has multiple sets of combinations of low pass filters, which receive a common pulse train signal, and parts for output-sampling. The signals that are obtained from the respective combinations of the low pass filters and the parts for output-sampling and that have different phases are combined and used. Thus, the number of times of sampling can be substantially increased and the accuracy of measurements can be increased.
The multiple sets of the combinations of the low pass filters and the parts for output-sampling may be at least two sets.
A common pulse train signal from the part for pulse-converting is entered into each low pass filter in each set. Each part for input-sampling samples the pulse train signal based on an input-sampling signal, and then the sampled signal is output. The input-sampling signals for the respective parts are synchronized and have different phases.
The parts for storing filter coefficients in the low pass filters in the respective sets read out the filter coefficients based on the respective readout signals. The readout signal in the same low pass filter is synchronized with and has the same phase as the input-sampling signals. The readout signals in the different low pass filters are synchronized with but have different phases from the input-sampling signals.
An adder and a subtracter in each set adds or subtracts filter coefficients read out from the part for storing filter coefficients in the set to or from the signal output from a part for storing results of an arithmetical process. Then the signals output from respective parts for output-sampling are added in time series and the result is output.
The pulse train signal is sampled by each input-sampling signal that is synchronized with, but has different phases from, the corresponding signals in the other sets. And each filter coefficient is read out based on each readout signal, which is synchronized with but has a different phase from the corresponding signals in the other sets. The respective convolutions between the pulse train signal and the filter coefficients are carried out. Thus output signals having different phases are obtained. The output signals are based on the signals that are sampled by signals having different phases. Therefore, the output signals are substantially equivalent to those obtained by a shorter sampling period.
The phase-difference may be, say, a time interval obtained by equally dividing a sampling period. For example, it may be a half of a sampling period, or 1/3 of a sampling period.

(Method for Measuring a Velocity)

A method for measuring a velocity comprises a step for converting, a step for low pass filtering, a step for output-sampling, and a step for outputting. In the step for converting, two-phase signals are converted to a ternary pulse train signal having a positive or a negative or zero value corresponding to the phases of the two-phase signals. In the step for low pass filtering, a low frequency component of the pulse train signal is obtained by means of arithmetical processing between the ternary pulse train signal and filter coefficients. In the step for output-sampling, the signal obtained by the step for low pass filtering is sampled at a predetermined sampling period. In the step for outputting, the sampled signal by the step for output-sampling is output as a velocity signal corresponding to the velocity determined by the two-phase signal.
In the step for low pass filtering, the pulse train signal is sampled at a predetermined period, and a convolution (operation for convolution) between the pulse train signal and the filter coefficients is carried out based on the sampled pulse train signal.
The convolution is a combination of adding and subtracting, in which a filter coefficient is added when the ternary pulse train signal is positive, and a filter coefficient is subtracted when the pulse train signal is negative. Reading the filter coefficients is synchronized with and has the same phase as sampling the pulse train signal.
In the step for output-sampling the result of the convolution is sampled at the sampling period and output as a velocity signal.
A method for measuring a velocity has multiple sets of combinations of a step for low pass filtering and a step for output-sampling. The step for low pass filtering uses a common pulse train signal. The outputs that have different phases and that are obtained by the respective combinations of the steps for low pass filtering and the step for output-sampling are combined. Thus, the number of times of sampling can be substantially increased and an output that is sampled by a shorter sampling period can be obtained.
In the sets of combinations of the step for low pass filtering and the step for output-sampling, for low pass filtering in the same set sampling the pulse train signal is synchronized with and has the same phase as reading the filter coefficients for the convolution. For filtering in different sets, sampling the signal is synchronized with but has a different phase from reading the coefficients.
The invention of this application can be applied not only to measuring a velocity of a spinning body but also to measuring a velocity of a linearly moving body.

Advantageous Effects of Invention

As above, the device for measuring a velocity of this invention does not include any integrating process. Therefore, the device can supress quantization noises caused by integrating quantization errors. Also, it can prevent the generation of aliases in a process for sampling.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is an illustration of an outline of a device for measuring a velocity of the invention of this application.

FIG. 2 is an illustration of the relationships between signals in a device for measuring a velocity of the invention of this application.

FIG. 3 is an illustration of an exemplary configuration of a device for measuring a velocity of the invention of this application.

FIG. 4 is an illustration of the relationships between signals in an exemplary configuration of a device for measuring a velocity of the invention of this application.

FIG. 5 is an illustration of a second exemplary configuration of a device for measuring a velocity of the invention of this application.

FIG. 6 is an illustration of a second exemplary configuration of a device for measuring a velocity of the invention of this application.

FIG. 7 is an illustration of an example for measuring a velocity in accordance with the invention of this application.

FIG. 8 is an illustration of an example for measuring a velocity in accordance with the invention of this application.

FIG. 9 is an illustration of an outline of a conventional device for measuring a velocity.

FIG. 10 is an illustration of relationships between signals in a conventional device for measuring a velocity.

DESCRIPTION OF EMBODIMENTS

Now, the embodiments of this invention are described in detail, with reference to the drawings. Below, an exemplary configuration of a device for measuring a velocity of this invention is described, with reference to FIGS. 1-8.
FIGS. 1 and 2 are illustrations of an outline of, and the relationships between signals in, a device for measuring a velocity of the invention of this application. FIGS. 3 and 4 illustrate the first exemplary configuration of a device for measuring a velocity of the invention of this application, and the relationships between the signals in the device. FIGS. 5 and 6 are illustrations of a second exemplary configuration of a device for measuring a velocity of the invention of this application. FIGS. 7 and 8 are illustrations of an example for measuring a velocity in accordance with the invention of this application.

(A Configuration of a Device for Measuring a Velocity)

Now, an outline of a device for measuring a velocity of the invention of this application is described, with reference to FIGS. 1 and 2. As in FIG. 1 a, in a device for measuring a velocity 1 of the invention of this application, an encoder 10 outputs two-phase encoded pulse train signals synchronized with a spinning body such as a rotating shaft. One of the two-phase encoded pulse train signals has a phase leading to, or lagging from, the other, depending on the direction of the movement, e.g., the direction of revolution. Based on the relationship between the phases, such information as the direction of the revolution can be obtained.
Part 2 for pulse-converting converts the signals to an up and down pulse train (+/− pulse train) signal that has three values, i.e., positive, negative, and zero. For example, the pulse train signal can be generated by detecting the edges of the two-phase signals. The sign, i.e., positive, negative, or zero, of the signal, can be obtained based on the change of the relationship between the phases.
FIG. 2 has illustrations explaining the pulse-converting. FIG. 2 a shows one example of encoded pulse trains having an A-phase and a B-phase, each of which has a different phase. There are four relationships between the A-phase and the B-phase. FIG. 2 a shows the relationships by the numbers 1-4.
The relationship between the A-phase and the B-phase has a different mode of shifting, depending on the direction of the movement. FIGS. 2 b and 2 c show the relationships corresponding to each direction of a revolution. FIG. 2 b shows the relationships of the phases when the spinning body is rotating clockwise. FIG. 2 c shows the relationships of the phases when the spinning body is rotating counterclockwise.
Part 2 for pulse-converting converts the encoded pulses having the A-phase and the B-phase to an up and down pulse train (+/−pulse train) signal that has three values, i.e., positive, negative, and zero. For example, when the relationship between the A-phase and the B-phase is changed from the respective phases of (0, 0) to those of (0, 1), a “+1” pulse is output as the pulse train signal. In contrast, when the relationship is changed from the respective phases of (0, 1) to those of (0, 0), a “−1” pulse is output as the pulse train signal.
FIG. 1 b shows encoded pulses and FIG. 1 c shows a pulse train signal (+/−pulse train) processed by the pulse-converting.
A low pass filter (LPF) 3 directly receives the pulse train signal that has been converted by part 2 for pulse-converting. It carries out an arithmetical process between the signal and filter coefficients and allows the low frequency component that is included in the pulse train signal to pass. Thus, it extracts a velocity signal component. Part 4 for output-sampling samples the signal output from the low pass filter 3 at a predetermined sampling period. The process for allowing low frequency components to pass through the low pass filter 3 and the process for sampling by the part 4 for output-sampling causes a velocity signal (FIG. 1 d) corresponding to the velocity indicated by the two-phase signal to be output.
FIGS. 1 e and 1 f illustrate how the reductions in the alias components differ by changing the procedures for low pass filtering and for sampling. FIG. 1 e illustrates a conventional process in which sampling is carried out before low pass filtering is completed. FIG. 1 f illustrates a process of the invention of this application in which low pass filtering is carried out before sampling is done. As in FIG. 1 e, if the low pass filtering is carried out after the sampling is done, the low pass filtering (LPF) (shown by a dashed line) cannot reduce alias components at a frequency (f_s-f_i) (shown by a dotted line) generated by the sampling.
In contrast, as in FIG. 1 f, if the sampling is carried out after the low pass filtering is done, since the low pass filter (LPF) (shown by a dashed line) reduces certain signal components, alias components at a frequency (f_s-f_i) (shown by a dotted line) generated by the subsequent sampling can be reduced.
(First Embodiment of the Device for Measuring a Velocity of the Invention of this Application)
Now, the first embodiment of the invention of this application is described, with reference to FIGS. 3 and 4. FIG. 3 illustrates the exemplary configuration of a low pass filter 3 that is included in a device for measuring a velocity. The exemplary configuration in FIG. 3 is a FIR filter.
The low pass filter 3 carries out a convolution (operation for convolution) between a pulse train signal (+/−pulse train) and filter coefficients. The filter coefficients can be, say, coefficients in a FIR filter.
The low pass filter 3 comprises a part 3 f for input-sampling, a counting part 3 e, a part 3 d for storing filter coefficients, a multiplexer 3 a, an adder 3 b, a subtracter 3 c, and a part 3 g for storing the results. The part 3 f for input-sampling samples the pulse train signal and takes the sampled signal. The counting part 3 e outputs a readout signal having a predetermined period, an input-sampling signal synchronized with and having the same phase as the readout signal, an output-sampling signal having a sampling signal, and a reset signal synchronized with the output-sampling signal, as timing signals for arithmetic processes, etc. The part 3 d for storing filter coefficients receives the readout signal from the counting part 3 e as a timing signal and sequentially reads out filter coefficients at a predetermined period, and outputs them. The multiplexer 3 a outputs one of three input signals using a pulse train signal as a control signal for a selection. The adder 3 b adds the filter coefficients read out from the part 3 d to the output from the part 3 g. The subtracter 3 c subtracts the filter coefficients read out from the part 3 d from the output from the part 3 g. The part 3 g stores the result of the convolution between the pulse train signal and the filter coefficients.
The counting part 3 e can generate a readout signal and an input-sampling signal having a predetermined period based on the clock signal from a part 5 for generating a clock signal. The part 3 e also can generate an output-sampling signal and a reset signal having a sampling period.
The multiplexer 3 a selects the output from the adder 3 b, i.e., the result of the addition, and outputs it, if the pulse train signal is positive. If the pulse train signal is negative, the multiplexer 3 a selects the output from the subtracter 3 c, i.e., the result of the subtraction, and outputs it. If the pulse train signal is zero, the multiplexer 3 a selects the value in the part for storing the results and outputs it. The part 4 for output-sampling receives the output-sampling signal as a timing signal and samples the output from the multiplexer 3 a. By sampling at the sampling period, the part 4 obtains the result of the arithmetic process for all the filter coefficients and then outputs a velocity signal.
The part 3 f for input-sampling receives the input-sampling signal as a timing signal, samples the pulse train signal having 2 values, and sends the result to a multiplexer 3 a. The part 3 d for storing filter coefficients reads out the filter coefficients that have been stored based on the readout signal, and sends them to the adder 3 b and the subtracter 3 c.
The input-sampling signal is synchronized with and has the same phase as the readout signal. Thus, sampling the pulse train signal and the convolution can be carried out with the same timing as the clock signal.
The counting part 3 e outputs a reset signal to a part 3 g for storing the results at a sampling period. The reset signal causes the part 3 g to start the convolution from the beginning of the sampling period and to finish it at the end of the period. In this way, the convolution is carried out during the sampling period, i.e., the period is a unit for the convolution. Thus, a velocity signal during each sampling period can be obtained.
FIG. 4 a is an exemplary illustration of a pulse train signal (+/−pulse train). A change from a positive (+1) value to a negative (−1) value shows that the direction of movement (the direction of revolution) has changed. FIG. 4 b is an illustration of sampling pulses that shows the period for measurements. By using those pulses the part for output-sampling carries out the sampling.
FIG. 4 c illustrates a clock signal. FIG. 4 d illustrates a readout signal and an input-sampling signal. The readout signal and the input-sampling signal can be formed based on the clock signal. For example, by counting a predetermined number of pulses in the clock signal, those signals can be formed.
FIG. 4 f illustrates a reset signal. FIG. 4 g illustrates an output-sampling signal. The reset signal and the output-sampling signal are formed based on the sampling period. The reset signal is output at the same time as the output-sampling signal. Thus, the result of the convolution, which has been carried out during the sampling period, can be output and then be reset.
FIG. 4 e illustrates the results of the addition or the subtraction in the convolution (operation for convolution). The convolution carries out adding the filter coefficients to the pulse train signal that exists within a sampling period or subtracting the filter coefficients from the pulse train signal that exists within a sampling period, based on whether the pulse train signal is positive or negative. In this example, the pulse train signal has one of two values, i.e., +1 or −1. The adder 3 b sequentially adds N+1 filter coefficients a₀-a_Nto the output from the multiplexer 3 a or the subtracter 3 c sequentially subtracts N+1 filter coefficients a₀-a_Nfrom the output from the multiplexer 3 a, and then the result within the corresponding sampling period is output.
FIG. 4 h illustrates the fact that the result of the convolution is output with the same timing as the output sampling, i.e., at the sampling period. The output value corresponds to the velocity signal.
(Second Embodiment of the Device for Measuring a Velocity of the Invention of this Application)
Now, the second embodiment of the invention of this application is described, with reference to FIGS. 5 and 6. The second exemplary configuration is composed of a plurality of combinations of low pass filters and parts for output-sampling. The outputs from the respective combinations having different phases are combined, and thus the number of times of sampling is substantially increased. Therefore, an output that is sampled by a shorter sampling period is obtained.
As in FIG. 5, the second exemplary configuration is composed of a plurality of parallel combinations of low pass filters 3 and parts 4 for output-sampling that are illustrated in the first exemplary configuration. The pulse train signal converted by the part 2 for pulse-converting is input to each combination of the filter 3 and the part 4.
FIG. 5 illustrates an example that is composed of three combinations of low pass filters 3 and parts 4 for output-sampling. Namely, it shows the combinations of a low pass filter 3A and a part 4A for output-sampling, a low pass filter 3B and a part 4B for output-sampling, and a low pass filter 3C and a part 4C for output-sampling.
In the respective low pass filters 3A, 3B, 3C, at the parts for input-sampling, which sample a pulse train signal, the phases of input-sampling signals are shifted so that the respective samplings have a different timing. Also, the phases of the readout signals are shifted so that the respective timings for reading-out differ from each other. Then the signals that are obtained by using those shifted phases are combined and thus the number of times for sampling is substantially increased. If it is composed of the three low pass filters 3A, 3B, 3C, the respective phases of the input-sampling signals and the respective phases of the readout signals can be shifted by 1/3 period from each other. The means for shifting the input-sampling signals and the readout signal can use, say, delay circuits.
The output from each part 4A, 4B, or 4C is input to a part 7 for arithmetical processing, and the part 7 outputs a velocity signal. For example, the part 7 sequentially outputs the outputs from the parts 4A, 4B, and 4C without modification. Or, the part 7 processes the outputs at each period for measurement in a time series and calculates the sum or the average of the outputs at each period.
Now, the third embodiment of the invention of this application is described, with reference to FIG. 6. The embodiment is composed of two parallel combinations of low pass filters 3 and parts 4 for output-sampling.
FIG. 6 illustrates an example that is composed of two parallel combinations, i.e., a combination of a low pass filter 3A and a part 4A for output-sampling and a combination of a low pass filter 3B and a part 4B for output-sampling.
The configuration of the filter 3A and the part 4A, and the low pass filter 3B and the part 4B, in FIG. 6, is almost similar to the configuration illustrated in FIG. 3. But the signals input to the respective counting parts 3 e differ.
The counting part 3 e of the low pass filter 3A receives a clock signal from a part 5 for generating a clock signal. Based on the period and the phase of the signal, the part 3 e generates a readout signal and an input-sampling signal having a predetermined period. Also, it generates a reset signal and an output-sampling signal having a predetermined period for measurement.
The counting part 3 e of the low pass filter 3B has a delay circuit 6. The delay circuit 6 shifts the signal in the counting part 3 e of the filter 3B. Based on the phase-shifted signal, a readout signal and an input-sampling signal are generated. Also, a reset signal and an output-sampling signal having a predetermined period for measurement are generated.
In accordance with the example of FIG. 6, the delay circuit 6 forms a phase difference by delaying the phase of the readout signal and the input-sampling signal by a half period.
(An Example for Measuring a Velocity of the Invention of this Application)
Now, an example for measuring a velocity of the invention of this application is described, with reference to FIGS. 7 and 8.
FIG. 7 a shows a pulse train output from an encoder using counted values. FIG. 7 b shows a pulse train signal (+/−pulse train) output from a part for pulse-converting. The positive parts of the pulse train signal correspond to the periods in which the counted values output from the encoder are increasing. The negative parts of the pulse train signal correspond to the periods in which the counted values output from the encoder are decreasing.
FIG. 7 c shows velocities measured by the invention of this application and velocities measured by a conventional method. FIG. 8 shows an enlarged view of FIG. 7 c.
FIG. 8 a shows an example of a velocity signal that can be obtained by the first embodiment of the invention of this application. FIG. 8 b shows an example of a velocity signal that can be obtained by a conventional way, i.e., by detecting the differences between sampled values. The dotted line shows the true velocities.
The example of the velocity signal in FIG. 8 shows that noise components caused by aliases, etc., can be significantly suppressed compared to noise components generated in a conventional device.
FIG. 7 d shows the error between the velocity at the input signal from the encoder and the output signal, from the measured velocity. B shows the error obtained by a conventional device for measuring a velocity. A shows an error obtained by a device for measuring a velocity by using the invention of this application. By the invention of this application, the error is found to be decreased.
FIG. 7 e shows the relationship between the error and the velocity. The error is between the velocity measured at the input signal from the encoder and the velocity of the output signal that is measured. The letter B shows the relationship between the error and the velocity, in which the measured velocity is obtained by a conventional device for measuring a velocity. The letter A shows the relationship between the error and the velocity, in which the measured velocity is obtained by a device for measuring a velocity by using the invention of this application. The horizontal axis of FIG. 7 e shows the velocity and the vertical axis shows the error. FIG. 7 e shows that by the conventional device for measuring a velocity, the errors between the input signal and the output signal exist independently from the velocity. In contrast, by the device for measuring a velocity by using the invention of this application, the errors between the input signal and the output signal only exist where the velocity is approximately zero and the direction of the movement is changed. The errors can be ignored when the velocity is not zero.
The invention of this application is not limited to the embodiments described above. It can be modified based on the philosophy of this invention, and the modified inventions should not be excluded from the scope of this invention.

INDUSTRIAL APPLICABILITY

The device for measuring a velocity of this application can be applied not only for measuring a revolving speed, but also for measuring a velocity of an object moving on a line or on a curve.

REFERENCE LIST

1 device for measuring a velocity
2 part for pulse-converting
3 low pass filter
3A C3B C3C low pass filter
3 a multiplexer
3 b adder
3 c subtracter
3 d part for storing filter coefficients
3 e counting part
3 f part for input-sampling
3 g part for storing the results
4 part for output-sampling
4A C4B C4C part for output-sampling
5 part for generating a clock signal
6 delay circuit
7 part for arithmetical processing
10 encoder
100 encoder
101 device for measuring a velocity
102 counting part
103 sampling part
104 part for detecting differentials

Claims

1. A device for measuring a velocity comprising:

a part for pulse-converting that converts two-phase signals to a ternary pulse train signal having a positive or a negative or a zero value corresponding to phases of the two-phase signals,

a low pass filter that outputs a low frequency component of the pulse train signal by means of arithmetical processing between the ternary pulse train signal and filter coefficients, and

a part for output-sampling that samples the signal output from the low pass filter at a predetermined sampling period and outputs a velocity signal corresponding to a velocity determined by the two-phase signal.

2. The device for measuring a velocity of claim 1, wherein

the arithmetical processing at the low pass filter is a convolution between the ternary pulse train signal and the filter coefficients,

wherein the low pass filter comprises:

a counting part that outputs, as timing signals, a readout signal having a predetermined period, an input sampling signal synchronized with the readout signal and having the same phase as that of the readout signal, and an output sampling signal having the sampling period,

a part for input-sampling that samples the ternary pulse train signal based on the input sampling signal,

a part for storing filter coefficients that stores the filter coefficients and that reads the respective filter coefficients responding to the readout signal,

a multiplexer that outputs one of three input signals by using the ternary pulse train signal as a signal for selection,

an adder that adds a filter coefficient read from the part for storing filter coefficients,

a subtracter that subtracts a filter coefficient read from the part for storing filter coefficients, and

a part for storing the results that stores the results of the convolution between the pulse train signal and the filter coefficients, wherein

the multiplexer selects a signal and outputs the selected signal from among the output signal from the adder, the output signal from the subtracter, and the output signal from the part for storing the results, wherein the multiplexer selects the output signal from the adder and outputs the output signal from the adder when the pulse train signal is positive, selects the output signal from the subtracter and outputs the output signal from the subtracter when the pulse train signal is negative, and selects and outputs the output signal from the part for storing the results when the pulse train signal is zero, and wherein

the part for storing the results samples the output signal from the multiplexer at the sampling period and outputs the sampled output signal from the multiplexer as a velocity signal.

3. The device for measuring a velocity of claim 2, wherein

the low pass filter comprises an FIR filter,

the counting part outputs a reset signal having the sampling period as one of the timing signals,

the adder adds the filter coefficient read from the part for storing filter coefficients to the output signal from the part for storing the results,

the subtracter subtracts the filter coefficient read from the part for storing filter coefficients from the output signal from the part for storing the results, and wherein

the part for storing the results resets the result of the convolution at the sampling period.

4. The device for measuring a velocity of claim 2, wherein

the device comprises at least two sets of combinations of the low pass filter and the part for output-sampling,

the low pass filter in each set receives a common ternary pulse train signal from the part for pulse-converting,

the part for input-sampling in each set samples the ternary pulse train signal based on the input-sampling signal that is synchronized with and has a phase different from those in the other low pass filters, and receives the sampled signal,

the part for storing filter coefficients in each set reads filter coefficients based on the read-out signal wherein the read-out signal is synchronized with and has the same phase as the input-sampling signal in the same low pass filter and wherein the readout signal is synchronized with and has a phase different from the input-sampling signals in the other low pass filters,

the adder and the subtracter in each set respectively adds and subtracts a filter coefficient read from the part for storing filter coefficients in each set, and wherein

the device arithmetically processes the output signals from the respective parts for output-sampling in sequence or in combination, and outputs the results.

5. A method for measuring a velocity comprising:

a step for converting in which two-phase signals are converted to a ternary pulse train signal having a positive or a negative or a zero value corresponding to phases of the two-phase signals,

a step for low pass filtering in which a low frequency component of the pulse train signal is obtained by means of arithmetical processing between the ternary pulse train signal and filter coefficients,

a step for output-sampling in which the signal obtained by the step for low pass filtering is sampled at a predetermined sampling period, and

a step for outputting in which the sampled signal by the step for output-sampling is output as a velocity signal corresponding to a velocity determined by the two-phase signal.

6. The method for measuring a velocity of claim 5, wherein

in the step for low pass filtering, the ternary pulse train signal is sampled at a predetermined period, a convolution between the pulse train signal and the filter coefficients is carried out based on the sampled ternary pulse train signal,

in the convolution, a filter coefficient is added when the sampled ternary pulse train signal is positive, and a filter coefficient is subtracted when the pulse train signal is negative, wherein reading the filter coefficients is synchronized with sampling the pulse train signal, and wherein

in the step for output-sampling the result of the convolution is sampled at the sampling period and output as a velocity signal.

7. The method for measuring a velocity of claim 5, wherein

the method has a plurality of sets of combinations of the step for low pass filtering and the step for output-sampling wherein a common pulse train signal is input in the respective sets of the combinations,

in the same step for low pass filtering, sampling the pulse train signal is synchronized with and has the same phase as reading the filter coefficients for the convolution, and wherein

between the steps for low pass filtering, sampling the pulse train signal in one of the steps is synchronized with and has a different phase from reading the filter coefficients for the convolution in any other of the steps.